Collaboration scripts in computer-supported collaborative learning: a systematic review from 2011 to 2024

Despite its crucial role in fostering learners' cognitive, social, and affective development, Computer-Supported Collaborative Learning (CSCL) continues to face persistent challenges, such as free-riding and unequal participation, that undermine its effectiveness. Collaboration scripts have been proposed as structured instructional supports that specify roles, sequence interaction, and provide procedural guidance to regulate collaborative processes. However, a comprehensive synthesis of recent empirical evidence remains limited. This study systematically reviews 53 peer-reviewed studies published between 2011 and 2024, examining collaboration scripts across four dimensions: research characteristics, learning outcomes, measurement approaches, and influencing factors. The findings indicate fluctuating publication trends, with renewed interest since 2020 driven by adaptive scripting and generative AI-supported approaches, reflecting a shift toward more flexible and learner-centered designs. Overall, collaboration scripts are associated with improvements across cognitive, behavioral, affective, and skills domains, though evidence regarding affective and skills-related outcomes remains mixed. Methodologically, studies rely on diverse instruments-including tests, questionnaires, user log data, artefacts, and interviews-yet single-method assessments are still common. Synthesizing influencing factors, this review conceptualizes collaboration scripts as regulation-distribution mechanisms whose impact emerges from the alignment among external supportive structures, learners' internal regulatory capacities, and contextual conditions. These findings provide theoretical and practical implications for the design and implementation of collaboration scripts in CSCL and suggest directions for future research.

10.17605/OSF.IO/VQFYC; https://osf.io/vqfyc/.